[gnoyeh525]

Originally Posted by hjasak

Hi Robert,

Thank you for looking into this. Can you please send me the files you have changed to review the bug fixes and merge them in.

I am looking into the best way to rewrite the parallel communications in the Immersed Boundary to make everything faster, but this is a bigger project so it will require real funding. The comms we have now are fine for a smaller number of processors, but as the decomposition increases, the code will become inefficient. I will speak to Zeljko to see what we should do; however, some significant funding is required to get this fixed.

Regarding the writing of U and other fields: at this point in the code, the data from the IB points is interpolated into the points of the STL surface. The algorithm I am using is a surface walk, which is linear if the STL is clean and does not protrude outside of the background mesh. Ideally, the resolution of the STL surface should be comparable with the resolution of the background mesh.

I assume you found a case where the surface walk algorithm fails, eg. if the STL is folded over or similarly broken.

Hrv

Dear Prof. Jasak

Is that possible to load in two or more STL geometries for immersed boundary ?
i.e. instead of 1 sphere, load in many spheres ... thanks as always

[hjasak]

Hi,

The STL can be in as many pieces as you like. I did not test the setup with multiple independent immersed boundaries in separate patches, so there may be problems related to coding; if you hit any we need some further development to deal with this.

You would only need to have true multiple IB patches (as opposed to having many bits of disconnected STL in a single immersed boundary) if you wish to post process eg forces on each bit separately.

Hope this helps,

Hrvoje

[Ben UWIHANGANYE]

Hello
May you help to plot drag and lift coefficient in movingcylinderinchannel tutorial/immersedboundary/foam-extent 4.0?

I have changed the controlDict like this but it is printing out Cd=0, Cl=0 and Cm=0 for all time steps. what could be the reason behind?.


functions
{
forces
{

type forceCoeffs;
functionObjectLibs ( "libforces.so" );
outputControl timeStep;
outputInterval 1;
patches
(
wall
);
pName p;
UName U;
rhoName rhoInf;
log true;
rhoInf 1;
CofR ( 0 0 0 );
liftDir (0 1 0);
dragDir ( 1 0 0 );
pitchAxis ( 0 0 0 );
magUInf 1;
lRef 1;
Aref 1;
}
}

[Nikrouz]

Hello everybody ....


I'm new to Foam extend .... as mentioned in release note of foam-extend 4.0, there have been significant improvements in Immersed Boundary Solvers. Is it possible to consider the immersed body as an "elastic structure"? If yes, which tutorials is suitable to start simulation?


Thanks everybody.

[zjz]

Originally Posted by hjasak

For the record: we will release the latest version of immersed boundary in foam-extend-3.2

Hrv

Dear Professor Jasak：
I have run the example of movingCylinderInChannelIco without change it in foam extend 4.0, but when i open it with paraview and run it, there is a fatal error "size of field refValue(96) is not the same size as the patch(0) on patch ibCylinder of field U in file"movingCylinerInChannelIco/0.2/U", what is the reason for this?

Thanks for your reply!

[mosaferkarbala]

Originally Posted by Phil_

Dear all,

so foam-extend next release (release 3.2) is available and I ran a small test for the immersed boundary.

For the kEpsilon turbulence model it works like a charm, but with the kOmegaSST model it complains about "tauWall not set for IB patch ibChannel for field U". In the kEpsilon model tauWall is set by the immersedBoundaryEpsilonWallFunction. This wall function is replaced by immersedBoundaryOmegaWallFunction where tauWall is not defined.

Any ideas?

In the meantime I'll dig in a little more and try to find out what's going on.

Phil

hey
do u find where this error comes from! im stuck on same error, any help would be appreciated
--> FOAM FATAL ERROR:
[0] tauWall not set for IB patch hull for field U
[0]
[0] From function const vectorField& immersedBoundaryVelocityWallFunctionFvPatchVectorF ield::wallShearStress() const
[0] in file wallFunctions/immersedBoundaryVelocityWallFunctions/immersedBoundaryVelocityWallFunctionFvPatchVectorF ield.C at line 200.

thanks

[Ginjames]

Originally Posted by Ben UWIHANGANYE

Hello
May you help to plot drag and lift coefficient in movingcylinderinchannel tutorial/immersedboundary/foam-extent 4.0?

I have changed the controlDict like this but it is printing out Cd=0, Cl=0 and Cm=0 for all time steps. what could be the reason behind?.


functions
{
forces
{

type forceCoeffs;
functionObjectLibs ( "libforces.so" );
outputControl timeStep;
outputInterval 1;
patches
(
wall
);
pName p;
UName U;
rhoName rhoInf;
log true;
rhoInf 1;
CofR ( 0 0 0 );
liftDir (0 1 0);
dragDir ( 1 0 0 );
pitchAxis ( 0 0 0 );
magUInf 1;
lRef 1;
Aref 1;
}
}

Have you solved this problem?I have a same one. please help me，thank you very much

[mosaferkarbala]

As u know there is no physical body in mesh because its immersed. So the wall type should be immersed boundary wall
Look at immersedboundary tutorials u will find it

[Ginjames]

I am so appreciated for your reply. But there is no tutorials in IBM of forcecoeffs. I'm still confused about the zero. Below is my controldict about force and forcecoeffs. The patch name is my immersed boundary. So what should I do？Looking forward to your help！
best wishes！

functions
(
forces
{
type immersedBoundaryForces;
functionObjectLibs ("libimmersedBoundaryForceFunctionObject.so");


outputControl timeStep;
outputInterval 1;
patches ( ibCylinder );

pName p;
UName U;
rhoName rhoInf;
rhoInf 1;

log true;
CofR ( 0 0 0 );

Aref 0.05;
Uref 1;
}

forces
{
type forceCoeffs;
functionObjectLibs ( "libforces.so" );
//functionObjectLibs ("libimmersedBoundaryForceFunctionObject.so");
outputControl timeStep;
outputInterval 1;
patches ( ibCylinder );
pName p;
UName U;
log true;
rhoName rhoInf;
rhoInf 1;
CofR ( 0 0 0 );
liftDir ( 0 1 0 );
dragDir ( 0 0 1 );
pitchAxis ( 0 0 0 );
magUInf 1;
lRef 2.83;
Aref 0.05;
}

[mosaferkarbala]

The first one in enough in foam extend immesrd bounday . Force coeff not needed.
After run afolder will create by name forces and u will find force and moments components inside it
Best regards

[Ginjames]

So the last question is that I want to get the force coefficient, such as drag of friction coefficient，to do some verification. What should I do？How to calculate？ Is the force in the forcefille in N/m3 or whatelse？ Best gratitudes for you. I have stumbled for a long time. thank you again.

[mosaferkarbala]

Originally Posted by Ginjames

So the last question is that I want to get the force coefficient, such as drag of friction coefficient，to do some verification. What should I do？How to calculate？ Is the force in the forcefille in N/m3 or whatelse？ Best gratitudes for you. I have stumbled for a long time. thank you again.

Hi dear brother
I also faced with that problem since the force coeffs always get zero since dosent sense the immersed body. The way i used computing drag using forces directly using drag coeff formula which is the force component in the motion direction and the surface
Best regards

[Ginjames]

So what is the dimension in the force fields？ I tried to calculate drag force coeffs，but can't agree with the literature. I think my formula is wrong. It’s a shame that I don't figure out the force file，is it in N/m3 or N/m or something else? How you calculate it？

[mosaferkarbala]

Originally Posted by Ginjames

So the last question is that I want to get the force coefficient, such as drag of friction coefficient，to do some verification. What should I do？How to calculate？ Is the force in the forcefille in N/m3 or whatelse？ Best gratitudes for you. I have stumbled for a long time. thank you again.

Originally Posted by Ginjames

So what is the dimension in the force fields？ I tried to calculate drag force coeffs，but can't agree with the literature. I think my formula is wrong. It’s a shame that I don't figure out the force file，is it in N/m3 or N/m or something else? How you calculate it？

Dimentions as i know is 3 major dimentions xyz, forces are in N and for drag
Fd =Cd*Ar*rhoV^2(1/2)
Fd is force in flow direction which in my case was -x direction, rho is ur fluid property and Ar is reference area which is 1D image of 3d shape that its surface vector is in flow direction
Thats it
For me this work quite fine
Best regards

[Ginjames]

Thank you brother. It’s very useful.

[cychen0318]

Hi,

I know this is an old post, but I am trying to implement the immerse boundary method in foam extend 4.0.

I have some problems with parallel running, although the error message is different from what you got before.
I use the same way to edit $WM_PROJECT_DIR/etc/controlDict and change commsType to nonBlocking as:

Originally Posted by Phicau

Hi Hua,


this error rings a bell to me:

[pjh-Veriton-D832:8321] *** An error occurred in MPI_Recv
[pjh-Veriton-D832:8321] *** on communicator MPI_COMM_WORLD
[pjh-Veriton-D832:8321] *** MPI_ERR_TRUNCATE: message truncated
[pjh-Veriton-D832:8321] *** MPI_ERRORS_ARE_FATAL: your MPI job will now abort

It was also triggered by a parallel boundary condition I developed. The solution was found thanks to Bernhard:

Edit $WM_PROJECT_DIR/etc/controlDict and change commsType to nonBlocking .

You should try and see if this helps with your problem.

Best,

Pablo

However, it seems that the problem is not solved and same error message pops up after I did the commsType change.


I have attached the log files from decomposePar and parallel run in the following:


This is the log for decomposePar:

// * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * //
Create time

Time = 0
Create mesh for region region0

Calculating distribution of cells
Selecting decompositionMethod simple

Finished decomposition in 0.61 s

Calculating original mesh data

Distributing cells to processors

Distributing faces to processors

Calculating processor boundary addressing

Distributing points to processors

Constructing processor meshes

Processor 0
Number of cells = 136711
Number of faces shared with processor 1 = 702
Number of faces shared with processor 2 = 11002
Number of processor patches = 2
Number of processor faces = 11704
Number of boundary faces = 1280

Processor 1
Number of cells = 136853
Number of faces shared with processor 0 = 702
Number of faces shared with processor 3 = 10961
Number of faces shared with processor 2 = 1
Number of processor patches = 3
Number of processor faces = 11664
Number of boundary faces = 1564

Processor 2
Number of cells = 136853
Number of faces shared with processor 0 = 11002
Number of faces shared with processor 3 = 681
Number of faces shared with processor 1 = 1
Number of processor patches = 3
Number of processor faces = 11684
Number of boundary faces = 1280

Processor 3
Number of cells = 136711
Number of faces shared with processor 1 = 10961
Number of faces shared with processor 2 = 681
Number of processor patches = 2
Number of processor faces = 11642
Number of boundary faces = 1564

Number of processor faces = 23347
Max number of processor patches = 3
Max number of faces between processors = 11704

Processor 0: field transfer
External flow
Number of IB cells: 100
triangles: 73264 hit: 100
Processor 1: field transfer
External flow
Number of IB cells: 0
triangles: 73264 hit: 0
Processor 2: field transfer
External flow
Number of IB cells: 100
triangles: 73264 hit: 100
Processor 3: field transfer
External flow
Number of IB cells: 0

and this is the log for parallel run:

Pstream initialized with:
nProcsSimpleSum : 0
commsType : nonBlocking
SigFpe : Enabling floating point exception trapping (FOAM_SIGFPE).

// * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * //
Create time

Create mesh for time = 0


PIMPLE: Operating solver in PISO mode

Create immersed boundary cell mask
Create immersed boundary face mask
Found immersed boundary patch 0 named ibSphere
External flow
[1] Number of IB cells: 0
[3] Number of IB cells: 0
[0] Number of IB cells: 100
[2] Number of IB cells: 100
Reading field p

Reading field U

Reading/calculating face flux field phi

Selecting incompressible transport model Newtonian
Selecting turbulence model type RASModel
Selecting RAS turbulence model kEpsilon
kEpsilonCoeffs
{
Cmu 0.09;
C1 1.44;
C2 1.92;
sigmaEps 1.3;
}


Starting time loop

Creating minMaxField for field p
Creating minMaxField for field U
Creating maxFieldCell for field epsilon
Time = 0.004

Courant Number mean: 0 max: 0.016 velocity magnitude: 0.25
PIMPLE: iteration 1
BiCGStab: Solving for Ux, Initial residual = 1, Final residual = 1.8743e-12, No Iterations 1
BiCGStab: Solving for Uy, Initial residual = 1, Final residual = 1.00661e-14, No Iterations 1
BiCGStab: Solving for Uz, Initial residual = 1, Final residual = 4.16333e-10, No Iterations 1
DICPCG: Solving for p, Initial residual = 1, Final residual = 0.0494606, No Iterations 132
IB time step continuity errors : sum local = 4.94896e-07, global = -8.45457e-09, cumulative = -8.45457e-09
DICPCG: Solving for p, Initial residual = 0.0150682, Final residual = 9.2989e-07, No Iterations 155
IB time step continuity errors : sum local = 1.04925e-09, global = 5.80027e-12, cumulative = -8.44877e-09
BiCGStab: Solving for epsilon, Initial residual = 0.0474431, Final residual = 1.09794e-11, No Iterations 1
BiCGStab: Solving for k, Initial residual = 0.309904, Final residual = 8.62141e-11, No Iterations 1
ExecutionTime = 11.25 s ClockTime = 11 s

triangles: 73264 hit: 100
[1]
[1]
[1] --> FOAM FATAL ERROR:
[1] tauWall not set for IB patch ibSphere for field U
[1]
[1] From function const vectorField& immersedBoundaryVelocityWallFunctionFvPatchVectorF ield::wallShearStress() const
[1] in file wallFunctions/immersedBoundaryVelocityWallFunctions/immersedBoundaryVelocityWallFunctionFvPatchVectorF ield.C at line 200.
[1]
FOAM parallel run aborting
[1]
--------------------------------------------------------------------------
MPI_ABORT was invoked on rank 1 in communicator MPI_COMM_WORLD
with errorcode 1.

NOTE: invoking MPI_ABORT causes Open MPI to kill all MPI processes.
You may or may not see output from other processes, depending on
exactly when Open MPI kills them.
--------------------------------------------------------------------------

Please let me know if you have encountered this issue or any idea how to solve this problem.
Thanks,


Cindy

[cookcaptain]

Hi,


Is it possible to output the skin friction on the immersed surfaces ?


Thanks

[hjasak]

Sure - use the wallShearStress tool. Immersed boundary is just a normal patch, although handled in a different ways.


Hrv

[cookcaptain]

Thanks a lot. The skin friction is extracted to the VTK file in each time step folder. But the VTK wall pressure has all zero values in the pitzDailyTurbulent case within foam-extend-4.1. It may be a bug for this particular case with the simpleFoam solver ?



Nevertheless the wall pressure can still be accessed from the usual volume solution by filtering the IB flag. The wall shear stress however does not work properly in the same way compared to the VTK data.



Moreover, the wall surface quantities are extracted on the near wall cell boundaries in foam-extend-4.1, instead of the STL surface in foam-extend-4.0. Is it for accuracy consideration ?




Shanggui

Originally Posted by hjasak

Sure - use the wallShearStress tool. Immersed boundary is just a normal patch, although handled in a different ways.


Hrv

[hjasak]

Not zero for me - are you sure you are looking at the right thing?

I run the tutorial to 150 iterations (not fully converged) and then wallShearStress. Attached is a picture: top is wall shear stress and bottom is pressure.

Hrv